Jan Behrens
Jan Behrens
After having given this topic some more consideration and after discussing this with a friend and colleague, I have doubts whether the current approach of named effects in Koka is...
> * You don't like how named effects have to be lexically passed I think my misunderstanding was that I thought of "named" effects as something like "named arguments", where...
> If I understand right, what you are proposing is something like: > > […] Almost, but not quite. ```diff -fun provide-config( action: () -> a ) : e a...
> […] when you just write `read()` what should be the inferred effect? `file`, `config` or `data`. […] Whichever effect is in scope, I guess? If there are no "neweffects"...
Maybe related to: * #769
Thanks for your explanations. I will think about the theoretical issues some more. I still feel like there are some issues with type-inference, and I also ran into an internal...
> This is related to the ordering of effect rows. Until now, I thought effect rows are some sort of "sets" where members can appear more than once though, but...